Antares rocket explosion: The question of using decades-old Soviet engines

Isn't this much more a question of reliability and track record than 10% performance differences?

how does the Vulcain2 compare?

The design is a technical marvel, but it won't keep the metal from corroding, and this isn't the first sign of trouble -- a few months ago, a different AJ-26 blew up on a test stand. It isn't yet clear that the engine was at fault in the Antares fire (trouble in the tankage or fuel lines could have cascaded down, among other fun possibilities) -- but if the engine does turn out to be at fault, the remaining stock will certainly get more attention.

What makes Rocketdyne engines less risky? Maybe they are both expensive and risky.

That's a pretty hard problem. F-1 took unimaginable efforts to make right - event though the only unusual, untested parameter for this engine is its size. RD-170 took all the genius of Glushko - including his previous somewhat successful experience with RD-270, which deserves a whole story for itself - to get working properly. With all hindsight, only the forth Merlin of SpaceX is actually pushing technical limits - and that's in XXI century. Then, there is the issue of time - to build an engine from scratch is easy to delay a rocket for a year - if you're lucky and do that right from the first attempt.

So Orbital words can be taken seriously and for the face value. NK-33 is a pretty good engine.

SpaceX certainly built their own and they seem to be working quite well thusfar.

Building a new engine is, quite literally, rocket science. On the plus side, we know the problems can be solved, because we've done it before. On the minus side, a lot of that knowledge has been lost and needs to be recreated or redesigned. We can't easily rebuild the Saturn V, for example, even though its design would work well for many of today's launches. [1] New rockets are being designed, but its a major undertaking. [2] NASA's SLS is in the $18 billion range. [3]

[1] http://amyshirateitel.com/2011/04/03/the-lost-art-of-the-sat...

[2] http://en.wikipedia.org/wiki/Saturn_V#Proposed_successors

[3] http://en.wikipedia.org/wiki/Space_Launch_System#Program_cos...

Minor nitpick: Frank Culbertson is an executive vice president at Orbital, not the CEO.

The CEO mentioned in the article (Elon Musk) said "there is better technology to use".

Those B-52's have been maintained (with some parts from boneyard-ed planes, though) and presumably at least occasionally flown all that time...

There's a lot less redundancy in a rocket. A B-52 can keep flying if a bunch of stuff stops working or falls off. You just cannot design a rocket that way.

"We saw rust and crack, so we repaired them" seems like the culprit. Not the age.

And the punchline is, those engines weren't even very good when they were first made all those decades ago.

If they article is right, they were first developed for use on the N1 rocket, the Soviet moon-rocket counterpart to America's Saturn V. There's a reason why Saturn V beat the N1 to the moon -- the N1 had a distressingly consistent habit of blowing up:

N1-L3 was underfunded and undertested, and started development in October 1965, almost four years after the Saturn V. The project was badly derailed by the death of its chief designer Sergei Korolev in 1966. Each of the four attempts to launch an N1 failed; during the second launch attempt the N1 rocket crashed back onto its launch pad shortly after liftoff and exploded, resulting in one of the largest artificial non-nuclear explosions in human history. (Source: http://en.wikipedia.org/wiki/N1_(rocket))

There hasn't been much development in this types of engines since the moon race. And we may have even gone backwards, because a lot of the institutional knowledge and probably some of the documentation is lost.

Yes and no.

Rocket engine design has not actually made that much progress in the meanwhile. It's not the design that was the issue.

Also, the Russians were VERY good at making rocket engines.

The real problem is that these are engines actually made decades ago, and stored under a tarp in a former Soviet facility for a long time before they were purchased by Orbital as the cheap option.

Context for the Musk Quotes, his 2012 Wired interview:

http://www.wired.com/2012/10/ff-elon-musk-qa/all/

Uninformed, yes :( .

There are theoretical limits for how high Isp - the specific impulse, defined as how much thrust, in, say, pounds of force, you can get from a unit of fuel flow, going into the engine, in, say, pounds of mass per second. The engines created in '60s are already rather close to that theoretical limit - so saying that you can do much better just because decades has passed is similar to saying "well, surely we had SR-71 flying back in '70s already - by today we ought to have all civil aviation routinely doing at least Mach 2".

So - no, it's pretty hard, including for fundamental reasons, to make engines much better that NK-33 by Isp. And that parameter is of paramount importance for a rocket engine.

Mind you, this is still not the only parameter. After all, you probably care a lot about economic efficiency. Elon Musk seems to understand that pretty well - and SpaceX engines, while lacking technically, could be in a much better position economically. Next, you're right that design and manufacturing aren't standing still. Both control of materials parameters and the ability to convert the raw material into an engineered device substantially rose - talking about, say, 3D printing engines or parts, which happens rather frequently now, but didn't in '70s.

So it's a complicated matter. However, NK-33 still have good technical parameters, and were bought inexpensively - in '90s, when Russia was severely starved for cash. Why not to use something which is already there, and not expensive? That's quite a rational decision Orbital - and Aerojet before them - took.

Regarding the second part, about storage. Surely this is unprecedented case. That's why the engines were tested - many times, both in Russia and in USA, and also analysed, trying to predict how several decades of storage could affect engines. The results were apparently positive - of course you and I don't have the original raw data which was used to make that decision. But don't think the engines just were, say, buried underground; their storage history is roughly known and isn't that different from other products of heavy machinery industry.

So I think Elon is habitually being overreacting in regard to Orbital usage of those engines. May be it has something to do with Elon's unsuccessful attempts to strike a deal with Russian space industry back when SpaceX was really young - or may be it's just a tradition for him, Elon sometimes says something out of proportions, even though SpaceX real - smaller - results are undeniable. Elon misses the point - it's only for uninformed the phrase "rocket engines made in USSR 40 years ago" would mean "unacceptable variant". George Sutton, author of "History of Liquid Propellant Rocket Engines" mentions that USSR spent far more towards rocket engine research than all other countries combined, and NK-33 is one of the serious achievements of Soviet rocket school. Don't discount something off hand just because it's old and foreign.

It's not that unique - in Russian space school oxidizer-rich turbines are routine. Proton, for example, used to have - until latest years - a pretty solid track record, especially for first stage engines, and those engines use N2O4 for turbines. Considering that N2O4 is burned with some amount of (CH3)2N-NH2, making HNO3... the gas going to chamber from turbine is the nitric acid. Still engines work just fine.

One of big reasons for Russians to use oxydizer-reach gas generators is that on the rocket there usually much more oxidizer - by mass - than fuel, which makes the flow through the turbine much bigger - several times. So efficiency of turbopumps goes up. That justifies the complexity of having oxidizer-rich gas flow before the chamber.

With these particular engines, you also have the variable factor of them being stored for decades in a warehouse, and all the issues with metal corrosion, fatigue, degradation, etc. that come with that.

I'm not a materials engineer, but I could imagine things like seasonal temperature changes, humidity, rough handling, changes in the metallurgical properties of the alloys they used over time could all affect the ability of the engine components to withstand high temperatures and pressures.

I think this is fundamentally pretty hard problem. It involves control of turbulence, and this is a big in general unsolved problem today. That includes hypersonic flights, thermonuclear plasma control... Rocket engines work on the edge of material capabilities, and fluctuations - which you want to tame to guarantee things like that won't happen - are hard to deal with.

Aerojet bought those engines from Russia - the original manufacturer - in 1990-s. No need for Orbital to scour museums.

No, engines didn't explode. Were actually never flown before the first Antares lifted off. But performed pretty well on stands.

No one is blaming Russia.

We're blaming the age of rocket engines (that just happened to come from Russia).

Not exactly. The engines themselves are decades old, Orbital Science is quoted as saying

"'As we went through testing, we did discover there were some effects of aging since they had been in storage for awhile, including some stress corrosion cracking,' Culbertson conceded at the time. 'That’s what we [corrected] with weld repairs and other inspections.'"

They got the engines that were decades old, and they've been fixing problems that they have found, and refurbishing them and installing them into rockets. It doesn't say they have built new engines on the old designs. That would have been better, because then you would have someone who had designed the engine on staff and could understand it in a deep way.

Presumably Orbital is designing their own engines now, or contracting with someone to do so, perhaps their go to market strategy was "Get some things into space with these old surplus engines, use the revenue/time to design/test/manufacture our own engines, and then continue on with our own bespoke design." I don't know of course if that was their intention or not.

The rocket engine is actually decades old. The engines were built many years ago, then stored under a tarp in a Russian facility somewhere. They were supposed to be dismantled, but some Party official risked his neck by sparing them the destruction, hiding them in some forgotten facility - however, the flip side to that was that the engines were stored in probably not the best conditions.

Orbital used them because they were the cheap option.

No, the engine is literally decades old. From http://en.wikipedia.org/wiki/NK-33 :

"Nearly thirty years after they were built, disbelieving rocket engineers were led to the warehouse. One of the engines was later taken to America, and the precise specification of the engine was demonstrated on a test stand."

No, in this case part of the rocket itself was decades old. Orbital used AJ-26 engines for the first stage, which are refurbished NK-33 engines originally built for the Soviet moon program. [1]

[1] http://en.wikipedia.org/wiki/NK-33

Actually, the engines are. The AJ-26 is a rebuilt NK-33, originally built in the early 1970s.

Downvote. The engines were indeed originally _built_ in the 60s or 70s and have literally been sitting around for the vast majority of the time since.

Second, not sure what your point is with the Russian launch. It uses different engines and was scheduled awhile back.

But cars are... they are extremely inefficient to more modern designs. We just haven't found something better then the internal combustion engine.

Not the rocket, the engines. And yes, they are decades old. The project that produced the NK-33 engines was abandoned i Soviet in 1974, so they are at least 40 years old. Aerojet bought a stockpile of them in the 90's.

The article refers to them as "refurbished Soviet engines" which certainly sounds like Some parts of the engines are original.